1. Hyperspectral Image Classification Bi-dimensional Empirical mode Decomposition and Deep Residual Networks
- Author
-
Ratnakar Dash, Ladi Sandeep Kumar, Ladi Pradeep Kumar, Harikiran Jonnadula, and Ganapati Panda
- Subjects
0209 industrial biotechnology ,Artificial neural network ,business.industry ,Computer science ,Feature extraction ,Hyperspectral imaging ,Pattern recognition ,02 engineering and technology ,Residual ,Hilbert–Huang transform ,020901 industrial engineering & automation ,Kernel (image processing) ,Computer Science::Computer Vision and Pattern Recognition ,Principal component analysis ,0202 electrical engineering, electronic engineering, information engineering ,Hyperspectral image classification ,020201 artificial intelligence & image processing ,Artificial intelligence ,business - Abstract
In this study a novel approach of hyperspectral image classification technique is realized using BEMD (Bi-Dimensional Empirical Mode Decomposition) and Deep Residual Networks. First Principal Component of the hyperspectral image dataset is computed using PCA(Principal Component Analysis) feature extraction technique. The model also adapts BEMD algorithm to divide the principle component into three hierarchical components and obtain BIMFs (Bi-Dimensional Intrinsic Mode Functions) and residue-image. These BIMFs and residue image is further taken as input to the deep residual network for classification. The results of experiments on two popular datasets of hyperspectral remote sensing scenes represent that the proposed-model offers a competitive analytical-performance in comparison to some established methods.
- Published
- 2020